WhatsApp Channel Join Now
Telegram Group Join Now

Relevance: GS-II (Government Policies, IT & Cyber Security); GS-III (Science & Technology, Artificial Intelligence) Source: MeitY / IT Amendment Rules, 2026

1 · What is the news in simple words?

Right now, we regulate Artificial Intelligence (AI) using bits and pieces of the old Information Technology (IT) Act, 2000. But the Ministry of Electronics and IT (MeitY) is now planning a brand new, standalone law just for AI!
Why? Because AI is no longer just a simple chatbot answering questions. We now have ‘Agentic AI’ (smart systems that can independently plan and execute tasks) and a massive rise in dangerous deepfakes! To protect citizens without stopping tech innovation, India is creating strict new rules for data consent, platform accountability, and ultra-fast deepfake removal.

2 · Why Do We Need a New AI Law?

Step 1: From Simple Prompts to ‘Agentic AI’
AI systems can now independently plan multi-step actions and use web tools without human supervision, requiring much tighter safety guardrails.
▼
Step 2: The Failure of Old ‘Safe Harbour’ Rules
Under old laws, social media platforms were immune because they just hosted user content. But since AI creates content, who is liable for harmful outputs?
▼
Step 3: Data Scraping & Consent Gray Areas
AI companies scrape our public social media photos and posts to train their models without clear consent, slipping past existing data protection laws!
▼
Step 4: MeitY’s Standalone Law Solution
India is drafting a dedicated law mandating digital watermarks, strict liability rules, and AI testing sandboxes with RBI and SEBI!

3 · Key AI Concepts

What is Agentic AI?
Autonomous Action Takers
Traditional AI only answers prompts. Agentic AI can independently plan tasks, use software tools, browse the web, and correct its own mistakes to achieve high-level goals with almost zero human supervision!
SGI & Deepfakes
Synthetically Generated Info
SGI is the legal term for deepfakes! It refers to audio, videos, or photos created by AI algorithms that look and sound so authentically human that ordinary citizens cannot tell they are fake.
Safe Harbour Erosion
Who is to Blame?
Section 79 of the IT Act gave tech platforms “safe harbour” (immunity) because they just acted as neutral postmen. But when generative AI actually writes defamatory text, the platform cannot easily claim immunity!
Regulatory Sandboxes
Safe Testing Zones
To avoid stifling tech startups, the government plans to set up regulatory sandboxes with bodies like RBI and SEBI. These are safe, controlled environments where companies can test new AI tools on real data without legal risk.

UPSC Prelims Quick Facts
IT Amendment Rules 2026 Mandates platforms to take down unlawful SGI (deepfakes) within 3 hours of a government notice! For severe violations like non-consensual intimate imagery, the takedown deadline is just 2 hours.
Digital Watermarking New rules force platforms to embed permanent metadata or unique digital markers into AI content so investigators can easily trace a deepfake back to its creator.
IndiaAI Mission Goals India is building domestic computing muscle! We are deploying over 100,000 GPUs via the IndiaAI Compute Portal and hosting culturally rich Indian datasets on the AIKosh platform.
DPDP Act 2023 Gap The Digital Personal Data Protection Act currently exempts personal data that a user voluntarily makes public, leaving our social media profiles open to AI scraping.

MCQ Practice Question
Q. With reference to Artificial Intelligence (AI) governance and digital regulations in India, consider the following statements:

  1. Unlike standard generative AI, ‘Agentic AI’ refers to autonomous systems capable of independently planning and executing multi-step workflows with minimal human supervision.
  2. Under the IT Amendment Rules, 2026, social media intermediaries are mandated to remove unlawful Synthetically Generated Information (SGI) within 24 hours of receiving an official notice.
  3. AIKosh is an indigenous dataset platform launched under the IndiaAI Mission to provide culturally representative data for training foundational AI models.

Which of the statements given above is/are correct?
(a) 1 and 2 only    (b) 1 and 3 only    (c) 2 and 3 only    (d) 1, 2 and 3

Answer: (b) 1 and 3 only

  • Statement 1 — Correct: Agentic AI represents the next frontier of AI, moving beyond static prompts to systems that autonomously plan actions, use web tools, and solve complex goals.
  • Statement 2 — Incorrect (the trap): Be careful with numbers in the exam! The 2026 IT Amendment Rules enforce an ultra-fast compliance window: unlawful SGI must be taken down within 3 hours (not 24 hours!), and within just 2 hours for intimate imagery.
  • Statement 3 — Correct: To prevent bias from Western datasets, the IndiaAI Mission established AIKosh to host high-quality, localized Indian datasets for training domestic AI models.
Ajmal IAS APSC Foundation Course

Recent Posts

Ajmal IAS APSC Foundation BATCH
Ajmal IAS APSC Foundation Course

Start Yours at Ajmal IAS – with Mentorship StrategyDisciplineClarityResults that Drives Success

Your dream deserves this moment — begin it here.